Why the Sky Has Green and Purple Stripes

That was my initial reaction to the paper “Why Liberals and Atheists are More Intelligent” by Satoshi Kanazawa.

While reading Satoshi Kanazawa's summaries of his alleged research, I noticed the following passages:

Analyses of large representative samples, from both the United States and the United Kingdom, confirm this prediction. In both countries, more intelligent children are more likely to grow up to be liberals than less intelligent children. For example, among the American sample, those who identify themselves as “very liberal” in early adulthood have a mean childhood IQ of 106.4, whereas those who identify themselves as “very conservative” in early adulthood have a mean childhood IQ of 94.8.

and

Once again, analyses of large representative samples from both the United States and the United Kingdom support this prediction of the Hypothesis. Net of a large number of social and demographic factors, including education, more intelligent individuals are more likely to be atheistic than less intelligent individuals. For example, among the American sample, those who identify themselves as “not at all religious” in early adulthood have a mean childhood IQ of 103.09, whereas those who identify themselves as “very religious” in early adulthood have a mean childhood IQ of 97.14.

Childhood IQ? The most obvious and most easily checked measurement would be of adult IQ. The obvious conclusion is that he did check the correlation with adult IQ but it didn't support his conclusion. (I won't more than mention the theory that all this means is that us wingnuts scored worse on a culturally-biased, meaningless test.)

I can think of two plausible reasons why liberalism etc. might be correlated with childhood IQ but not with adult IQ:

• Liberals might grow up faster. Anecdotal evidence indicates they stop being virgins earlier. On the other hand, similar evidence says they reproduce later.
• People who think that more of society should be run by civil service might be briefed earlier on the civil-service way of looking at the world.

Since there are reasons for high-IQ people to at least pretend to be liberals (and vice versa), it's possible that us wingnuts have higher IQs. According to that same theory, we would expect that us high-IQ wingnuts are more likely to be complete nerds. (I am, of course, speaking for myself.) There is some unreliable, anecdotal evidence to back up that claim.

There are other criticisms of the hypothesis.

A Digression

The above reasoning is based on the following heuristic:

1. Figure out what Joe Bloggs (an average reader) would conclude from the report. If the report was strongly stated, it was probably either written by an activist who was trying to get people to believe that conclusion or by someone who based it on the activists' press releases. (In this case, Joe Bloggs would conclude that atheist liberals are the real intellectuals.)
2. Determine the strongest potential piece of evidence that would point in the same direction. If that evidence were true, the report would have mentioned it. (In this case, it would be reports of a correlation with adult IQ.)
3. In the absence of such evidence being mentioned, conclude that it doesn't exist.

(The above heuristic will also enable you to conclude that if a food is labeled “cholesterol-free,” it is probably high in fat unless the label says otherwise.)